Cellular immune responses to recurring influenza strains have limited boosting ability and limited cross-reactivity to other strains

Transcription

1 ORIGINAL ARTICLE VIROLOGY Cellular immune responses to recurring influenza strains have limited boosting ability and limited cross-reactivity to other strains Y. Keynan 1, C. M. Card 1, B. T. Ball 1,2,4,Y.Li 2, F. A. Plummer 1,2,3 and K. R. Fowke 1,3,4 1) Department of Medical Microbiology, University of Manitoba, Manitoba, Canada, 2) Public Health Agency of Canada, Winnipeg, Manitoba, Canada, 3) Department of Community Health Sciences and 4) Department of Medical Microbiology, University of Nairobi, Nairobi, Kenya Abstract Influenza vaccine provides protection against infection with matched strains, and this protection correlates with serum antibody titres. In addition to antibodies, influenza-specific CD8+ T-lymphocyte responses are important in decreasing disease severity and facilitating viral clearance. Because this response is directed at internal, relatively conserved antigens, it affords some cross-protection within a given subtype of influenza virus. With the possibility of a broader A(H1N1) Mexico outbreak in the fall of 2009, it appeared worthwhile studying the degree of cellular immune response-mediated cross-reactivity among influenza virus isolates. The composition of the influenza vaccine included the A/New Caledonia/20/1999 strain (comprising a virus that has been circulating, and was included in vaccine preparations, for 6 7 years) and two strains not previously included (Wisconsin and Malaysia). This combination afforded us the opportunity to determine the degree of cross-reactive cellular immunity after exposure to new viral strains. We analysed the antibody responses and the phenotype and function of the T cell response to vaccine components. The results obtained show that antibody responses to A/New-Caledonia were already high and vaccination did not increase antibody or cytotoxic T lymphocyte responses. These data suggest that repeated exposure to the same influenza stain results in limited boosting of humoral and cellular immune responses. Keywords: CD8+ T cell, cellular immunity, HIA, influenza, intracellular staining, memory subset, vaccination Original Submission: 7 July 2009; Revised Submission: 7 September 2009; Accepted: 15 December 2009 Editor: E. Gould Article published online: 23 December 2009 Clin Microbiol Infect 2010; 16: /j x Corresponding author and reprint requests: K. R. Fowke, Department of Medical Microbiology, University of Manitoba, Room Bannatyne Avenue, Winnipeg, Manitoba, R3E 0J9, Canada fowkekr@cc.umanitoba.ca Introduction Efficacy of influenza vaccines has been correlated with a rise in serum antibody titres, as long as the virus strain used in the vaccine resembles the strain in circulation. After influenza vaccination, the humoral immune response requires several days to effectively contain viral spread and eradicate the virus. The immunologic memory established in this primary response results in long-lived resistance to re-infection with homologous virus. Cross-protection within a subtype of influenza has only rarely been observed and infections induce essentially no protection across subtypes or between types A and B [1,2]. Anti-haemagglutinin (HA) antibodies protect against both disease and infection with homologous virus. Serum HA-inhibiting titres of 1:40 or greater confer protection against infection [3]. In recognition of these defined immune correlates, the induction of neutralizing antibodies is one of the main goals of immunization. CD8+ cytotoxic T lymphocytes (CTL) recognize epitopes of HA or internal proteins M, NP, or PB2 presented in the context of HLA class I molecules [2]. Depending on their antigen specificity, CTLs may be subtype-specific; however, many target relatively conserved, internal antigens and are thus broadly cross-reactive among influenza A strains [4,5]. This extensive cross-protection has attracted recent attention in light of the swine-origin influenza A (H1N1) outbreak as well as the fact that 60% of patients are under 18 years of age and that questions exist about the degree of cross-protection conferred by A (H1N1) and other influenza strains [6,7] Cytokine responses may contribute to the elimination of influenza-infected cells; however, it is considered that CD8+ effector T cells [8] eradicate infected cells mainly by CD95 (Fas)-mediated or perforin- and granzyme-mediated cytotoxicity [9]. Intracellular cytokine staining profiles indicate that Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases

2 1180 Clinical Microbiology and Infection, Volume 16 Number 8, August 2010 CMI influenza epitope specific CD8+ T cells are capable of producing interferon (IFN)-c, often in combination with tumor necrosis factor and occasionally interleukin (IL)-2 [10]. Measures of cellular immune response to influenza are correlated with protection against influenza in the absence of strong serum Ab responses among the elderly, and combining cellular and humoral measures of vaccine efficacy may increase the ability to predict the risk of influenza illness [11,12]. Crossreactive T cell responses target relatively conserved internal proteins and may confer some protection against novel emerging strains. Lee et al. [13] demonstrated recognition of multiple H5N1 peptides, predominantly from viral matrix and nucleoprotein, by individuals residing in the UK who were not likely exposed to H5N1 virus. The study suggests that regular boosting of heterosubtypic CTLs may provide a broad protection against avian and human influenza [13,14]. After initial viral encounter, CD8+ memory T cells are characterized by reduced requirements for co-stimulatory signals in comparison to naïve T cells and a quicker response to antigenic restimulation [15 17]. The memory subsets are differentiated based on surface expression of lymph node homing chemokine receptor CCR7 and CD45RA. Phenotypically, cells co-expressing CCR7 and CD45RA are naïve, whereas CCR7+ CD45RA are central memory and CCR7 CD45RA are effector memory. Because CTL responses are directed toward relatively conserved epitopes, each re-exposure in adults is a secondary response to some extent. McElhaney et al. [18] compared single-dose and two-dose influenza immunization in older adults and documented a ten-fold decline in IL-10 levels after the first dose. This response persisted at least 16 weeks after vaccination. The study revealed a correlation between the IFN-c:IL-10 ratio and antibody titres in the single-dose group and a negative correlation in the two-dose group. They concluded that a Th1-mediated antibody response to the initial vaccination and a Th2 response to the boosting, two-dose regimen accounted for the difference and that only the Th1 response elicited by single-dose vaccination would lead to effective CTL cytokine production [18]. Inactivated influenza vaccine is the most common vaccine used, with an efficacy of 70 90% for preventing infection with influenza virus, and with the correlate of protection being a rise in serum antibody titre. Ohmit et al. [19] documented a four-fold or greater antibody titre rise from pre-vaccination levels among 66.7% of inactivated vaccine recipients. Although these titres explain the protection afforded by vaccine to young healthy adults, levels are typically lower among elderly individuals with underlying medical conditions as well as when the circulating viruses do not match the strains included in the vaccine [20,21]. The recommended trivalent influenza vaccine virus strains included A/New Caledonia/20/1999 (H1N1)-like, A/ Wisconsin/67/2005 (H3N2)-like and B/Malaysia/2506/2004- like antigens [19]. One of these three strains, A/New Caledonia/20/1999, has been circulating and included in vaccine preparations for the last 6 7 years (and thus is referred to as old strain in the text). This provides a unique opportunity to study immune responses to a recurring influenza vaccine strain and to compare these with the responses to novel flu vaccine strains. We conducted a study to determine antibody and CTL response to the three vaccine strains in healthy adults. We specifically sought to analyse the contribution of CD8+ T cell memory subsets, referred to as central memory (T CM ) and effector memory (T EM ), after influenza vaccination. We also aimed to explore the role of immune exhaustion by repeated vaccination with recurring vaccine components through comparison between the old vaccine strain and the two new stains [A/Wisconsin/67/2005 (H3N2) and B/Malaysia/2506/ 2004] included in the vaccine preparation for the first time during the season in Winnipeg, Canada. Materials and Methods Study groups Study subjects included 14 healthy adult donors. The mean age of participants was 37.5 years (range years). Seven of the volunteers had received influenza vaccine in preceding years and seven had not. The volunteers attended a vaccination clinic at the University of Manitoba, Winnipeg, Canada. All subjects provided their informed consent, and the study was approved by the ethics review committee of the University of Manitoba. Participants were immunized with inactivated trivalent influenza vaccine according to the Northern Hemisphere recommendation for Blood for baseline and follow-up sampling was obtained on days 7 and 30. Peripheral blood mononucleocytes (PBMCs) were isolated from heparinized whole blood by density gradient centrifugation using Ficoll-hypaque (Bio-Lynx, Brockville, Canada) and washed twice in RPMI media (Sigma- Aldrich, St Louis, MO, USA) containing 10% FBS (Gibco, Carlsbad, CA, USA). Clinical laboratory tests Plasma was separated by centrifugation and stored at )80 C. Determination of antibody titres by a haemagglutination inhibition assay (HIA) was performed at the National Influenza Reference Laboratory, National Microbiology Laboratory, Public Health Agency of Canada, Winnipeg.

3 CMI Keynan et al. Cellular immune responses to recurring influenza 1181 Phenotyping and functional flow cytometry Following PBMC isolation, cells were stained using a panel of fluorochrome-conjugated mab including: CD4-AmCyan, CD3-Alexa Fluor 700, CD8-Pacific Blue, CD62L-PE, CCR7- PE-Cy7, HLA DR-APC-Cy7 and CD38-APC (all from BD Biosciences, San Diego, CA, USA) and CD45RA-ECD (Beckman Coulter, Fullerton, CA, USA). Cells were then washed in Dulbecco s phosphate-buffered saline (PBS) containing 2% foetal calf serum (FCS) (Gibco) and resuspended in PBS. One hundred thousand events per sample were acquired using an LSRII flow cytometer and analysis was performed with FACSDiva software (BD Biosciences). Following separation, PBMCs from each donor were incubated in the presence of media (negative control), staphylococcal enterotoxin B (SEB; Sigma), phytohaemagglutinin (positive controls), and each of the three influenza vaccine strains namely: A/New Caledonia/20/1999 (H1N1) (NC), A/Wisconsin/67/2005 (H3N2) (WISC) and B/Malaysia/2506/ 2004 (MAL). The antigens were added at a final concentration of 10 haemagglutinin units/ml. The influenza strains were obtained from the National Influenza Reference Laboratory, Winnipeg, Canada. After 3 days of in vitro culture, cells were incubated for a further 6 h in the presence of GolgiPlug (BD Biosciences) to prevent secretion of cytokines, followed by staining for flow cytometric analysis. After antigen stimulation, PBMCs were surface-stained using a panel of fluorochrome-conjugated mab including: CD4-AmCyan, CD3-Alexa Fluor 700, CD8-Pacific Blue, CD62L-PE, CCR7-PE-Cy7, HLA DR-APC-Cy7 (all from BD Pharmingen, San Diego, CA, USA) and CD45RA-ECD (Beckman Coulter). Cells were then washed in Dulbecco s PBS containing 2% FCS (Gibco) and intracellular staining was performed. After surface staining, intracellular cytokine staining was performed using the BD Cytofix/Cytoperm Fixation/Permeabilization Kit according to manufacturer s instructions (BD Biosciences). Cytokines were detected using IFN-c-fluorescein isothiocyanate and IL-2-allophycocyanin (BD Biosciences). One hundred thousand events per sample were acquired using an LSRII flow cytometer and analysis was performed with FACSDiva software (BD Biosciences). The percentage of cytokine-positive cells within the parent population was determined and the stimulation index was calculated as the ratio of cytokine-positive cells in stimulated vs. unstimulated cultures. Statistical analysis Mean values (percent cells expressing a particular phenotype) were compared between groups using the nonparametric Mann Whitney U-test; p 0.05 was considered statistically significant. The Kruskal Wallis nonparametric test was used for the calculation of trend between the three time points. Results Antibody titres HIA was performed on all samples for each of the three vaccine components. Mean baseline titres were 52 (95% CI ), (95% CI 95 )6.45 to 31.26) and (95% CI ) for New Caledonia/20/1999 (H1N1), A/ Wisconsin/67/2005 (H3N2) and B/Malaysia/2506/2004, respectively. HIA titres on day 0 (before vaccination) and days 7 and 30 (after vaccination) for each of the strains contained in the vaccine are shown in Fig. 1. Over this time period, HIA titres for NC did not change, but statistically significant increases in antibody titres for WISC and MAL were noted (Kruskal Wallis p values for trend were: NC, 0.16; WISC, and MAL, ). Cellular immunity Functional response to antigen-specific stimulation after vaccination. To determine the function of CD8+ T cells in response to vaccine components, we stimulated freshly isolated PBMCs with whole influenza viruses corresponding to the strains included in the vaccine, at timepoints prior to vaccination, and 7 and 30 days after vaccination. After separate incubation with each of the three vaccine strains, intracellular cytokine staining was performed as previously described [22]. In vitro stimulation with influenza virus was used to describe the cytokine production in response to antigenic stimuli p = n.s p = p = Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 New-Caledonia Wisconsin Malaysia FIG. 1. Haemagglutination inhibition assay titres for each of the vaccine components were determined for each study participants (New Caledonia, Wisconsin, Malaysia) before vaccination (day 0) (checkered bar) and at two time points after vaccination (days 7 and 30) (black and grey bars respectively); p values represent the trend between three time points using Kruskal Wallis analysis of variance. NS, not significant.

4 1182 Clinical Microbiology and Infection, Volume 16 Number 8, August 2010 CMI Assessment of antigen-specific functional responses (IL-2 and IFN-c) is shown in Fig. 2. Briefly, we gated on lymphocytes and, subsequently, on CD3+ and CD8+ cells. Within the CD8+ cells, we gated on IFN-c and IL-2. This response was compared with that of the same individual to NC (the old vaccine component). We found increased IL-2 and IFNc production by CD8+ T-cells stimulated in vitro with WISC compared to NC. Phenotype and state of immune activation. Memory subset classification was performed by gating on CD8, and dividing the CD8+ T cells into naïve CD8+ CD45RA+, central memory (T CM ) CD8+ CD45RA) CCR7+ and effector memory (T EM ) cells CD8+ CD45RA) CCR7). Memory subsets were determined just before and 7 days and 30 days after vaccination with inactivated influenza virus. The distribution of memory subsets within the entire CD8 T-cell population did not change significantly after vaccine administration. This likely reflects that influenza-specific responses represent a small portion of the total T cell repertoire. Immune activation, as assessed by CD38 expression, was measured for each of the three memory subsets (naïve, T CM and T EM ). The proportions of activated CD8+ T cells were determined and shown to be increased significantly after immunization with inactivated influenza virus in the naïve subpoulation only (Fig. 3). Figure 4 depicts the data for all three vaccine strains together. The IFN-c production by CD8+ T EM cells increased significantly after immunization, whereas the proportion of CD8+ T CM cells that produce IFN-c showed a significantly decreased response. This likely represents the differentiation of central memory T cells into effector and effector memory cells. Comparison of the functional response to new vs. old strains after vaccination. To determine the function of influenza CD8+ T cells to repeated vaccination, we compared the IFN-c and IL-2 production in response to stimulation with old and new vaccine components, prior to and after vaccination with trivalent inactivated vaccine. (a) Specimen_ DAY 7 (b) Specimen_ DAY 7 SSC-A (x 1000) P1 Count P2 (c) CD8 Pacific Blue-A Specimen_ DAY 7 CD8 CD FSC-A (x 1000) CD3 AmCyan-A CD4 Alexa700-A (d) 45RA ECD-A Specimen_ DAY 7 P4 P6 P5 (e) IL-2 APC-A Specimen_ MEDIA Q1-2 Q2-2 IL2 POS IFNG POS Q3-2 Q CCR7 PeCy7-A IFNG FITC-A FIG. 2. Gating strategy. The gating strategy employed was to first gate on lymphocytes (a) followed by CD3+ cells (b) and finally T cell subsets (CD4 or CD8) (c). Then memory subsets, defined by CD45RA and CCR7, were gated on (d). Functional responses on either T cell subsets (CD4/CD8) or memory subsets (naïve, T CM,T EM ) were assessed by determining interleukin (IL)-2 and interferon (IFN)-c responses (e) using the described intracellular staining protocol.

5 CMI Keynan et al. Cellular immune responses to recurring influenza 1183 % activated p = NS NS 0 Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 Naive T EM T CM *Kruskal-Wallis test FIG. 3. Total activation state within the CD8+ memory subsets. The activation state is expressed as the percentage of CD8+ T cells within the population expressing the activation surface marker CD38; p values represent the trend between three time points using Kruskal Wallis analysis of variance. NS, not significant. FIG. 4. Interferon (IFN)-c response within the CD8+ memory subsets. The responses pre- and post-vaccination to all three vaccine components are plotted together; p values represent the trend between three time points using Kruskal Wallis analysis of variance. % of IFN producing CD P = 0.01 IFNγ production within memory subsets P < Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 T EM T CM Gated on IL-2 producing cells, the memory subset of responding T cells was determined using expression of CD45RA and CCR7. The IL-2 produced by T EM in response to new components increased significantly after immunization, whereas IL-2 produced by T CM decreased. Vaccination did not significantly alter the IL-2 production within either subset in response to stimulation with the old strain (Fig. 5). The same procedure was repeated for IFN-c. Gating on IFN-c-producing cells, the memory subset of responding T-cells was determined using expression of CD45RA and CCR7. The IFN-c produced by T EM in response to new components increased significantly after immunization, whereas no such increase was observed in response to the old virus (Fig. 6). Discussion The influenza vaccine composition as recommended by the WHO for season in the Northern Hemisphere presented an opportunity to assess the degree of % CD8 T cells p = New strains IL-2 new versus old Old strain p = p = NS Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 T EM T CM T EM T CM FIG. 5. Strain-specific interleukin (IL)-2 responses within memory subsets of new strains (Wisconsin and Malaysia) compared to the old strain (New Caledonia). The responses are expressed as the percentage of CD8+ T cells of that memory class (either T EM or T CM ); p values represent the trend between three time points using Kruskal Wallis analysis of variance. NS, not significant.

6 1184 Clinical Microbiology and Infection, Volume 16 Number 8, August 2010 CMI % of IFNγ Producing CD p = New virus IFNγ Effector memory CD8 cross-reactive cellular immunity afforded by a circulating virus that is present in vaccine formulations (NC) and to compare it with that afforded by the new components of the vaccine (WISC and MAL). The present study was designed to carefully assess the the phenotype and function of T cell responses to the new and old vaccine components, with particular focus on the T cell memory compartments. The antibody response measured using HIA showed higher pre-immunization titres against the old strain; however, the titre remained unchanged after vaccination. The new strains started with lower HIA titres and then significantly increased after 7 and 30 days. The higher baseline titre for the old strain and lack of increase after vaccination was observed even in the group that did not receive the vaccine in previous years, suggesting that the circulating virus may have contributed to the already protective antibody at the outset (data not presented). Influenza-specific CD8+ T cells aid in the efficient elimination of virus and in mediating subsequent host recovery. These tasks are accomplished via the production of proinflammatory cytokines and the direct killing of virus-infected cells [4,5]. Some of the T cell responses to influenza virus are directed at evolutionarily conserved sites within different strains and are important for cross-strain reactivity, a notion that provided the impetus for the present study. The cytokine production within memory subsets, in response to immunization, was studied using matching strains of wholevirus as a T cell stimulus, and functional measurements consisted of intracellular staining for IL-2 and IFN-c. The proportion of CD8+ IFN-c+ lymphocytes within the effector memory subset increased, whereas it decreased within the central memory compartment. These reciprocal changes suggest that, after vaccination, the T CM differentiate to become IFN-c-producing effectors resulting in a decrease in central memory cell frequency and an increase in the effector memory NS Old virus 0 Day 0 Day 7 Day 30 Day 0 Day 7 Day 30 FIG. 6. Strain-specific interferon (IFN)-c responses within memory subsets, with new strains (Wisconsin and Malaysia) compared to the old strain (New Caledonia). The responses are expressed as the percentage of CD8+ T cells of that memory class (either T EM or T CM ); p values represent the trend between three time points using Kruskal Wallis analysis of variance. NS, not significant. cell type. Studying the reponse to specific epitopes using tetramer staining is needed in the future for a better understanding of the clonal nature of the expanding T CM subset. IL-2 and IFN-c production by CD8+ T-cells increased significantly 30 days after influenza vaccination in response to new strains but not the old strain (Figs 5 and 6). The IFNc response to the latter strain occurred earlier, peaking after 7 days, followed by a decrease on day 30. One possible explanation for the decreased CD8+ T cell response to the old influenza strain could be the presence of higher baseline neutralizing antibody titres, attenuating the secondary T cell response by limiting the antigen availability for presentation to T cells. In the presence of less neutralizing antibody, more antigen is available to induce an anamnestic T cell response. This mechanism was shown to be operating in the case of influenza infection [23,24]. These results are in contrast with the well-demonstrated effect after exposure to a virus in which the memory cell pool expands, leading to higher precursor frequency upon secondary exposure. The expansion occurs at the expense of nonreacting memory T-cells [25]. A conceivable explanation for the discrepancy is that repeated exposure to a circulating strain with its variants leads to expansion of different clones directed at re-assorted internal proteins, such that these replace pre-existing cells by a process of attrition.functional exhaustion of the CD8+ T cells has been reported mostly in chronic infections, characterized by persistence of antigens such as hepatitis C virus and HIV [26,27]. Repeated or high doses of antigen may be associated with induction of tolerance or depletion of responding CTLs. Using a vaccinia-virus epitope with variable expression levels, Wherry et al. [28] demonstrated that the highest expression was associated with a larger memory pool; however, the function of these cells, as measured by IFN-c, was diminished. Although it is not considered to be important in acute transient viral infections, we speculate that recurring exposure to NC circulating for several years along with repeated vaccination could have led to dysfunctional memory and hence the poor cytokine production by CD8+ T cells observed in the present study. The CD8+ T cell responses are considered to be important with respect to protection from heterologous viruses and, because they target preserved viral epitopes, they hold promise for the prevention of potential epidemic strains of influenza virus. These cross-reactive T cells can become activated and modulate the immune response and the outcome of subsequent heterologous infections, and they have been shown to expand at the cost of loss of noncross-reactive clones [29,30]. Skewing of the CTL cytokine response toward a Th2 phenotype upon repeated vaccine administration [18] could

7 CMI Keynan et al. Cellular immune responses to recurring influenza 1185 explain the lack of increase in IFN-c production observed with NC. On the other hand, the absence of boosting of the antibody response is in contrast to observations made by McElhaney et al. [18]. The observation of decreased cellular response to repeated administration of a component of the influenza vaccine, namely NC, compared to new strains is important. The implications are that such re-vaccination with the same strain may limit the development of a CD8 response and might decrease the potential protection against heterologous influenza virus, which is considered to be mediated to a large extent by these cells. The data also suggest that vaccination with multiple influenza isolates is required to develop the breadth in T cell response. The data may also be relevant when considering potential vaccines against the H1N1 Mexico influenza virus. Our data suggest that, although limited cross-protective cellular immunity is offered by exposure to other influenza strains, the maximal breadth and specificity of T cell responses is only achievable with strain-specific immunization. Studies that directly address the cross-protective immunity conferred by H1N1 Mexico and circulating strains are needed. The present study has demonstrated three notable points. The first is the observation of the evolution over time of a memory subset of responding cells, documenting the conversion of central memory T cells to a larger effector memory compartment after vaccination. The second is the lack of boosting of the antibody and CTL response by re-administration of a vaccine component that has been circulating and included in the vaccine formulation for several years. The final point is that isolate-specific immunization is required for a maximal breadth of T cell responses. This comprehensive study of both antibodies and CD8+ T cells with comparison of the three vaccine constituents provides information that may be useful for future recommendations of vaccine formulations. Transparency Declaration Funding was received from Bill and Mellinda Gates, Challenges in Global Health, Public Health Agency of Canada. YK is a fellow of the CIHR International Infectious Disease and Global Health Training Program. CMC holds a PhD graduate scholarship from the Canadian Institutes for Health Research (CIHR). FAP holds a Canada Research Chair in Resistance and Susceptibility to Infections. KRF is the recipient of a New Investigator award from the CIHR and holds a Manitoba Research Chair from the Manitoba Health Research Council. The authors declare that they have no conflicts of interest. References 1. Tamura S, Tanimoto T, Kurata T. Mechanisms of broad cross-protection provided by influenza virus infection and their application to vaccines. Jpn J Infect Dis 2005; 58: Doherty PC, Turner SJ, Webby RG, Thomas PG. Influenza and the challenge for immunology. Nat Immunol 2006; 7: Hibson D. Cost effectiveness of influenza vaccination. J. Hyg. (Lond.) 1972; 70: McMichael AJ, Gotch FM, Noble GR, Beare PA. Cytotoxic T-cell immunity to influenza. N Engl J Med 1983; 309: Doherty PC, Topham DJ, Tripp RA. Establishment and persistence of virus-specific CD4+ and CD8+ T cell memory. Immunol Rev 1996; 150: Novel Swine-Origin Influenza A (H1N1) Virus Investigation Team. Emergence of a novel swine-origin influenza A (H1N1) virus in humans. N Engl J Med : Fraser C, Donnelly CA, Cauchemez S et al. Pandemic potential of a strain of Influenza A (H1N1): early findings. Science : Marshall DR, Olivas E, Andreansky S et al. Effector CD8+ T cells recovered from an influenza pneumonia differentiate to a state of focused gene expression. Proc Natl Acad Sci USA 2005; 102: Topham DJ, Tripp RA, Doherty PC. CD8+ T cells clear influenza virus by perforin or Fas-dependent processes. J Immunol 1997; 159: La Gruta NL, Turner SJ, Doherty PC. Hierarchies in cytokine expression profiles for acute and resolving influenza virus-specific CD8+ T cell responses: correlation of cytokine profile and TCR avidity. J Immunol 2004; 172: McElhaney JE, Xie D, Hager WD et al. T cell responses are better correlates of vaccine protection in the elderly. J Immunol 2006; 176: McElhaney JE, Herre JM, Lawson ML, Cole SK, Burke BL, Hooton JW. Effect of congestive heart failure on humoral and ex vivo cellular immune responses to influenza vaccination in older adults. Vaccine 2004; 6: Lee LY-H, Hado LA, Simmons G et al. Memory T cells established by seasonal human influenza A infection cross react with avian influenza A (H5N1) in healthy individuals. J Clin Invest 2008; 118: Doherty PC, Kelso A. Toward a broadly protective influenza vaccine. J Clin Invest 2008; 118: Woodland DL, Scott I. T cell memory in the lung airways. Proc Am Thorac Soc 2005; 2: Ely KH, Cookenham T, Roberts AD, Woodland DL. Memory T cell populations in the lung airways are maintained by continual recruitment. J Immunol 2006; 176: Woodland DL. Cell-mediated immunity to respiratory virus infections. Curr Opin Immunol 2003; 15: McElhaney JE, Hooton JW, Hooton N, Bleackley RC. Comparison of single versus booster dose of influenza vaccination on humoral and cellular immune responses in older adults. Vaccine 2005; 23: Ohmit SE, Victor JC, Rotthoff JR et al. Prevention of antigenically drifted influenza by inactivated and live attenuated vaccines. N Engl J Med 2006; 355: Jackson LA, Jackson ML, Nelson JC, Neuzil KM, Weiss NS. Evidence of bias in estimates of influenza vaccine effectiveness in seniors. Int J Epidemiol 2006; 35: Skowronski DM, Masaro C, Kwindt TL et al. Estimating vaccine effectiveness against laboratory-confirmed influenza using a sentinel physician network: results from the season of dual

8 1186 Clinical Microbiology and Infection, Volume 16 Number 8, August 2010 CMI A and B vaccine mismatch in Canada. Vaccine 2007; 25: Betts MR, Brenchley JM, Price DA et al. Sensitive and viable identification of antigen-specific CD8+ T cells by a flow cytometric assay for degranulation. J Immunol Methods 2003; 2: Recommended composition of influenza virus vaccines for use in the influenza season Wkly Epidemiol Rec. 2006;81: Greenspan N, Doherty PC. Modification of cytotoxic T-cell response patterns by administration of hemagglutinin-specific monoclonal antibodies to mice infected with influenza A viruses. Hybridoma 1982; 1: Greenspan N, Hurwitz JL, Doherty PC. Capacity of influenza virusmonoclonal antibody mixtures to stimulate memory and cytotoxic T lymphocyte populations. Infect Immun 1982; 38: Kostense S, Vanderberghe K, Joling J et al. Persistent numbers of tetramer+ CD8(+) T cells, but loss of interferon-gamma+ HIV-specific T cells during progression to AIDS. Blood 2002; 99: Lechner F, Wong DK, Dunbar PR et al. Analysis of successful immune responses in persons infected with hepatitis C virus. J Exp Med 2000; 191: Wherry EJ, McElhaugh MJ, Eisnlohr LC. Generation of CD8(+) T cell memory in response to low, high, and excessive levels of epitope. J Immunol 2002; 168: Kim SK, Welsh RM. Comprehensive early and lasting loss of memory CD8 T cells and functional memory during acute and persistent viral infections. J Immunol 2004; 172: Klenerman P, Zinkernagel RM. Original antigenic sin impairs cytotoxic T lymphocyte responses to viruses bearing variant epitopes. Nature 1998; 394: